AI Signal 142
Anthropic job listing seeks specialist to track activists as threats; firm has reported users to police but refused to share messages
↗
A Prospect investigation, citing Anthropic job postings and a podcast with the firm's Global Security Operations Center, describes Anthropic building predictive activist-monitoring and reporting users to police without sharing the underlying messages.
For engineers whose products handle user speech, the precedent set here is that in-platform statements can be escalated to law enforcement without the reporting company disclosing what was actually said, making the conversation itself opaque to the person being reported. For anyone working in or near AI policy, the gap between Anthropic's public stance against domestic surveillance and the threat categories its security team is hiring to monitor is the concrete tension worth watching.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
An Anthropic job posting for an enterprise intelligence specialist at $180,000, $230,000 lists activism alongside terrorism and nation-state targeting as a category to be tracked.
Anthropic contracts with Samdesk for risk detection and used its protest-intelligence to reroute an executive away from a planned demonstration with about 60 minutes of advance notice.
The San Francisco Standard reported Anthropic called police on a Claude user over an in-platform statement, then refused to share the actual messages with officers, citing internal policy.
THE CLUSTER